A Kalman filter-based stable dynamic inversion for discrete-time, linear, time-varying systems
| Authors: | Iftime Orest, Delft Center for Systems and Control, TU Delft and Faculty of Economics, University of Groningen, Netherlands
Verhaegen Michel, Delft Center for Systems and Control, TU Delft, Netherlands |
|---|
| Topic: | 1.4 Stochastic Systems |
|---|
| Session: | Linear and Nonlinear Filtering in Stochastic Systems |
|---|
| Keywords: | Discrete-time systems, Time-varying systems, Input estimation, Inversion, Kalman filter |
|---|
Abstract
In this paper the problem of estimating an unknown input fordiscrete-time, non-minimum phase, multivariable, lineartime-varying systems (LTV) is considered. The initial condition ofthe plant may be unknown and stochastic process and measurementnoise are included. The input signal is modelled as a random walkwith drifts. Then it is estimated using a Kalman filter for auniformly detectable augmented system. A necessary and sufficientcondition for the detectability of the augmented system isprovided. A Kalman filter-based stable dynamic inversion forLTV systems is obtained as a consequence of our solution to theproposed problem. The inversion technique can be applied toachieve output tracking. We are mainly motivated by typical need to replicate time signals in the automobile industry.
